Two additional species of Gymnopus (Euagarics, Basidiomycotina)

Abstract For more than a decade, a combination of molecular phylogenetic analyses and morphological characterisation has led to a renovation of the Omphalotaceae, especially of Gymnopus sensu lato. Numerous new genera have been proposed, but Gymnopus sensu stricto has also seen an accretion of species and species complexes. In this manuscript, two species are added to Gymnopus sensu stricto within Section Androsacei.


Introduction
Ongoing research on marasmioid and gymnopoid fungi (Antonín and Noordeloos 2010;Antonín et al. 2014;Mata et al. 2004;Hughes 2016, 2017;Wilson and Desjardin 2005), has led to significant renovation of Gymnopus sensu lato. Several additional genera have been proposed and molecular phylogenetic analyses have revealed numerous small clades within Gymnopus sensu stricto. One such clade includes Marasmius brevipes Berk. & Ravenel. The result is the necessary transfer of M. brevipes to Gymnopus and proposal of a new species, G. portoricensis.
Nomenclaturally, recombination of Marasmius brevipes into Gymnopus produces a conflict between two potential homonyms, of which Gymnopus brevipes (Bull.) S.F. Gray has priority. A new name is required for Marasmius (Gymnopus) brevipes and this is introduced below as Gymnopus neobrevipes.

Materials and methods
The following abbreviations and acronyms are noted: RHP, KWH = initials of the authors; GSMNP, Great Smoky Mountains National Park; M = Marasmius; Ma = Marasmiellus; Mi = Micromphale; My = Mycetinis. Colour names enclosed in quotation marks ("") are from Ridgway (1912) and those cited alphanumerically are from Kornerup and Wanscher (1967). BF = bright field microscopy; PhC = phase contrast microscopy. Microscopic structures were observed in 3% aqueous potassium hydroxide (KOH) without staining. Spore metrics are expressed as Q = the range of spore length divided by spore width; Q m = mean value of Q.
All photos of microscopic structures were taken using a Q c Olympus camera mounted on an Olympus BX60 research microscope fitted with phase contrast microscopy.
Molecular methods were described in Petersen and Hughes (Petersen and Hughes 2016; see also Petersen and Hughes 2017). An LSU-based PhyML phylogeny illustrates general placement of section Androsacei within Gymnopus and related taxa ( Fig. 1). An ITS-based PhyML phylogeny was constructed to show more detailed placement of the two species below within Section Androsacei (Fig. 2). ITS and LSU sequences used in this paper are available in GenBank. Aligned ITS and LSU sequences are available in the Dryad depository (ITS: https://doi.org/10.5061/dryad.rd1df0c; LSU : https://doi.org/10.5061/dryad.4081h).
Habitat and phenology. Basidiomata on dead small branches of broad-leafed trees, in temperate forests often on fallen branches of Quercus or Rhododendron maximum in mixed forest including Tsuga, usually at or near ground level; sterile rhizomorphs decumbent on dead, small (usually 18-24 mm diam.) boughs. In tropical climates, (see Pegler 1983, 1987, 1988, Dennis 1953, 1970) encountered year-round; in temperate forests mid-Summer to early Autumn. Gymnopus neobrevipes sometimes  shares the same habitat as Anthracophyllum lateritium (Berk. & Curtis) Singer -dead Rhododendron maximum boughs over streams.
Commentary. Although collected by Ravenel, it was Curtis who conveyed the type specimen to Berkeley and Berkeley is the name-giver. The protologue (assumedly  "Pileus 1-2 line broad, convex, dark blood red; margin even; stem filiform, jet black, quite smooth, 1-2 line high, springing from creeping mycelial thread of the same nature with itself; gills ventricose, few, adnate, rufous.
"Allied to M. haematocephalus, &s, but distinguished at once by its short polished stem and dark gills. The colour of the pileus is nearly that of M. atrorubens. " The pileipellis structure is similar to others described in sect. Androsacei. Desjardin and Petersen (1989) described pileipellis as not gelatinised (the tissue is not so), but failed to describe the gelatinisation of individual hyphae. This gelatinisation is merely a minor gelatinous sheath of individual hyphae for the outline of hyphal wall is not altered. However, the flake-like encrustation is carried on the gel surface and is seen at some small distance from the hyphal wall outline.
Amongst basidia in a mount soaked in KOH overnight, structures are seen which can be interpreted as gelatinised cheilocystidia. In rare cases, the remnants of digitate branching can be seen, but usually nothing is left of the supporting cell but some ghost-like structure. In a mount of lamellar edge only briefly in KOH, an enormous amount of debris is detected surrounding hymenial structures. It appears to be some sort of degeneration, quite possibly partial gelatinisation, but including numerous rodshaped bacteria. This may be another indication of gelatinisation of tissues, this time of old basidia and subhymenial hyphae.
Subbasidial hyphae (subhymenium) become zig-zag in form as basidia are formed, evacuate and disappear. These hyphae are easily mistaken for some sort of cystidial structures, especially cheilocystidia.
A chronology of authoritative literature follows. Singer (in Dennis 1953) supplied a detailed description of G. neobrevipes (as Mi. brevipes) and Dennis (1953) examined type material and offered a rather uninformative illustration. Dennis (1970) offered a description of M. brevipes (as Micromphale), but perhaps as valuable is a diminutive aquarelle (Pl. 8, Fig. 4) which provides a good representation. Perhaps the best description of G. neobrevipes (as Mi. brevipes) was offered by Pegler (1983) and the description was based on more specimens than the type. Desjardin (1989: 447-449) examined the type of M. brevipes and Desjardin and Petersen (1989) published a species description based on numerous specimens.
Pileipellis structures, especially erect broom cell-like cells, are often gelatinised, especially in age. Likewise, cheilocystidia, while observed only occasionally, are often reduced to debris by gelatinisation or occasionally produce apical growths which can attain significant length. Lamellar tramal hyphae are often observed as thick-walled, but usually this is due to gelatinisation of the hyphal walls (inner wall boundary is clear, but outer wall boundary is obliterated and the gelatinised wall appears as though thick). Pegler (1983) included Mi. brevipes as the only representative of Micromphale in the Lesser Antilles. Rhizomorphs of G. neobrevipes are viable and short surface-sterilised sections (circum 1 cm) placed on malt extract agar produce sprays of mycelium from severed ends within 24 hrs. Within 72 h, the emergent mycelial sprays can be excised to establish an independent dikaryon culture which can be used for sequencing. In the case of G. neobrevipes, not only are sprays of mycelium produced on the cut ends (Fig. 4C), but within 72 h, many lateral hyphae emerge from the rhizomorph surface, soon resembling brownish fur. In one case, rhizomorphs of an ambient air-dried specimen were stored for over a year, yet produced mycelium as noted. In another case, a collection was heat-dried, but over one month later, rhizomorphs remained viable.
In Desjardin and Petersen (1989), a comparison was made of M. brevipes Berkeley & Ravenel, (1853;type, South Carolina, K) to Marasmius porphyreticus Petch (1947; type, Sri Lanka, K). They concluded that M. porphyreticus "differs mainly in absence of cheilocystidia and in forming 'plicatosulcate' pilei." Other discriminating characters of M. porphyreticus: "slightly thinner pileus context, more regularly forked lamellae, and basidiomata not arising directly from rhizomorphs." Significantly, presence or absence of clamp connections was not mentioned. Plicatosulcate pilei, forked lamellae and basidiomata arising from rhizomorphs are all found in G. neobrevipes as well as other similar basidiomata. The comparison, despite a disparate geographic distribution, remains questionable, possibly pending phylogenetic data.
Commentary. Although basidiomata superficially resemble those of G. neobrevipes, the pileipellis structure is not similar. Erect, broom cell-like cells of G. neobrevipes are missing; diverticulate repent hyphae are rare and doubtful; erect "hairs," while clamped (and therefore assumed to belong to this organism), are more demonstrable in G. neobrevipes. Morphologically, G. portoricensis could be placed in Marasmiellus (see Retnowati 2018) based on poorly developed Ramealis-structure, no broom cells), but it equally could be interpreted as a reduced member of Androsacei (including G. neobrevipes) in which erect, broom cell-like pileipellis cells are rare to missing. Cheilocystidia are typical of the latter group. If G. neobrevipes is accommodated in Gymnopus sect. Androsacei, G. portoricensis must also be found there. ITS sequences confirm this placement (Fig. 2).
Inspection shows that almost no basidiomata originate from rhizomorphs, instead seemingly originating from woody substrate directly. Rare basidiomata, however, do arise from rhizomorphs, with stipes as side branches. Moreover, some twigs with basidiomata are devoid of rhizomorphs altogether. A polyspore dikaryon culture was established from TENN-F-050999 and careful examination revealed exceedingly rare (but clearly demonstrated) clamp connections. This condition is also true in cultures of G. neobrevipes. Desjardin (1990), while reporting clamp connections in the culture of M. brevipes, made no comment on their relative abundance. Basidiomata are not pseudo-or eccentrically stipitate, but centrally to slightly eccentrically stipitate. The stipe, however, is usually immediately curved through the declivity in the pileus circumference. Lamellae appear to deteriorate rapidly, perhaps through insect grazing or tissue gelatinisation, but when discrete are shallow but sharp- ly defined (not merely as folds). Interlamellar anastomoses are absent and even lamellar buttressing is missing. Instead, the interlamellar hymenophore is smooth.
These two collections fruited on very different substrata. The origin within bamboo structures would be difficult to imagine, so perhaps basidiomata arise from a very thin, arachnoid mycelium on the bamboo surface. Rare basidiomata were seen attached to rhizomorphs, which might support typical attachment to somatic hyphae.
If G. portoricensis is regarded as in Marasmius, the epithet (portoricensis) is preoccupied by Marasmius portoricensis Murrill in Pennington. 1915. North American Flora 9(4): 262. The homonym is in Marasmius but not in Gymnopus. Described as having the longest ("longissimus") stipe -6-8 cm × 0.5 mm -and pileus 4-10 mm broad, the holotype of Marasmius portoricensis is at NY (isotype MICH) and the Mycoportal record shows several long-stiped basidiomata with stipe yellow-orange and apparently several long, straight rhizomorphs of similar colour.

Discussion
Singer (1948) proposed Micromphale sect. Rhizomorphigena based, in part, on his perception of gelatinisation of tissues in the pileus of the type species, Marasmius westii Murrill, (1945). Desjardin (1989) and Desjardin and Petersen (1989) concluded that diagnostic characters of M. brevipes matched those of M. westii and nomenclaturally, the epithet brevipes took priority. Moreover, these same characters more closely resembled those of Marasmius sect. Androsacei Kühner (1933) than those of Micromphale and they transferred Singer's section as Marasmius sect. Rhizomorphigena. César et al. (2018) considered Marasmius brevipes and M. westii as taxonomic synonyms and transferred the latter as Gymnopus westii. Based on our current examination of the type specimen of M. westii (FLAS-F-17211), we reject this synonymy. Some differences: 1) Hymenial elements are without clamp connections in M. westii while clamp connections are common in all tissues in M. brevipes (Desjardin and Petersen 1989 and this study); 2) pleurocystidia are not mammilate; 3) rhizomorphs are considerably thinner than those of G. neobrevipes; and 4) Murrill's notes with the type of M. westii describe rhizomorphs as "aerial" (i.e. suspended above ground level) while those of G. neobrevipes are at or near ground level, predominantly bound to fallen substrate with some aerial elements.